Modeling and Satellite Remote Sensing of the Meteorological Impacts of Irrigation

In the summer of 2012, the Central Plains of the United States experienced one of its most severe droughts on record. In order to study both the drought and human response to the drought (irrigation), land surface temperature (LST), Normalized Difference Vegetation Index (NDVI), and cloud fraction data from the Moderate-resolution Imaging Spectroradiometer (MODIS) are analyzed and a parameterization of irrigation using the Community Land Model (CLM) 4.0 coupled to the Weather Research and Forecasting (WRF) model (version 3.6.1) is created. According to the MODIS data, non-irrigated areas often show 5 K LST increases and negative NDVI anomalies (compared to summer 2002-2011 averages) while irrigated areas show < 2 K LST anomalies and NDVI anomalies near zero. As expected, the cloud fraction anomaly is negative nearly everywhere in the domain. However, the largest reduction in cloud fraction is found over the heavily-irrigated area, which conflicts with previous modeling studies showing an increase in cloud fraction over irrigated areas. This could be explained by a hypothesis that the temperature gradient between irrigated and non-irrigated areas is strong enough during severe drought to produce a local circulation (similar to land/sea breezes) that results in an atmospheric downdraft over the irrigated area. To investigate this process further, a parameterization of irrigation is developed in the CLM 4.0 coupled to the WRF model. The irrigation parameterization includes the following: (1) accurate representation of the spatial distribution of irrigation in the domain; (2) addition of soil moisture to irrigated areas; and (3) use of crop modeling to calculate accurate leaf area index (LAI) values throughout the growing season in both irrigated and non-irrigated areas. The modification of the default spatial distribution of irrigation in WRF, combined with LAI values calculated via crop modeling, account for two aspects of irrigation modeling that have often been neglected in previous work.